Speech, Hearing and Language: work in progress Volume 13 Individual differences in phonetic perception by adult cochlear implant users: effects of sensitivity to /d/-/t/ on word recognition

This study examined whether the phonetic perceptual phenomena associated with categorical perception in normal-hearing listeners (i.e., sharp identification functions, poor within-category sensitivity, high between-category sensitivity) are predictive of individual differences in speech recognition performance among cochlear implant patients. Adult postlingually deafened cochlear implant users, who were heterogeneous in terms of their implants and processing strategies, were tested on 2 phonetic perception tasks using a synthetic /s@/-/c@/ continuum (phoneme identification and discrimination) and 2 speech recognition tasks using natural recordings from 10 talkers (open-set word recognition and forced-choice /t/-/d/ recognition). Cochlear implant users tended to have identification boundaries and sensitivity peaks at voice onset times (VOT) that were higher than found for normalhearing individuals. Sensitivity peak locations correlated with individual differences in cochlear implant performance; individuals who had a /t/-/d/ sensitivity peak near normal-hearing peak locations were most accurate at recognizing natural recordings of words and syllables. However, speech recognition was not strongly related to identification boundary locations or to overall levels of discrimination performance. The results suggest that at least a subset of the perceptual phenomena associated with categorical perception have a functional role in word recognition by cochlear implant users. Introduction The ability of individuals to recognize speech via cochlear implants calls for a reconsideration of what types of phonetic information and perceptual processing are necessary for human speech recognition. Cochlear implants bypass much of the auditory periphery, such that the neural firing patterns resulting from cochlear implant stimulation differ from normal neural firing patterns (e.g., Rubenstein et al., 1999). The functional number of frequency channels is fewer than for normal hearing (e.g., Dorman et al, 2000; Fishman et al., 1997), but temporal resolution can be about the same (e.g., Busby et al, 1993; Shannon, 1989, 1992; see Shannon, 1993 for a review). Despite the facts that cochlear implant stimulation is quite different from normal hearing in many respects, and that the standard frequency-related phonetic cues (e.g., formant and burst frequencies) may be difficult to discern given the poor spectral resolution of cochlear implants (e.g., Dorman, 1991; Shannon et al., 1995), the best users of current cochlear implants are able to recognize more than 90% words correct in clinical tests of open-set sentence recognition (e.g., Parkinson et al., 1998). This word recognition ability is particularly remarkable for postlingually deafened cochlear implant users. Adults are better at recognizing speech via cochlear implants if they had acoustic hearing during childhood (e.g., Naito et al., 1997; Tong et al., 1988; Waltzman et al., 1992), even though their electrical stimulation differs from the acoustic stimulation that guided their phonetic and lexical development. The normal-hearing perceptual phenomena associated with the categorical perception of consonants (e.g., sharp identification boundaries, low within-category sensitivity, and high between-category sensitivity; Liberman et al., 1957; Studdert-Kennedy et al.,